In a multimedia database system that captures and stores multimedia data such as video and audio in accordance with some event, it is known that there is currently no real time automated mechanism for intelligently controlling sensors (e.g., cameras), that is, the selecting and deselecting of sensors, and the setting of sensor parameters to steer the sensors to an object or person of interest, based on reasoning about the continual activity of objects or people in the event.
In the broadcast of sports, for instance, a large crew of trained camera personnel manually control various cameras, while a director continually chooses camera streams and orders switches between the different cameras.
There are security systems (such as BehaviorTrack, which is part of Loronix Video Solutions available from Verint Systems Inc. of Woodbury, N.Y.) that move cameras to preset locations upon a security breach. However, this camera movement is based on preset locations and cannot continually move the sensor to focus on a moving object or person. Furthermore, the decision to steer the camera to a preset location cannot be based on reasoning about the continual activity of objects or people in an event.
There are existing systems that control cameras using microphones. Typically, such technology is used in teleconferencing in order to focus the camera on the current speaker. However, the application of such an approach to other sensors in a multimedia database system that captures and stores multimedia data associated with a real time event is quite limited.
Thus, there exists a need for techniques that overcome the above-mentioned drawbacks by enabling the intelligent and automatic selection and control of sensors associated with a multimedia database system, which captures and stores multimedia data associated with a real time event.